inertia tensor

The inertia tensor is a compact bundle of numbers that describes how hard it is to start, stop, or turn a rigid body's rotation — and crucially, how that difficulty changes depending on which axis you spin it about. For straight-line motion, a single number (the mass) tells you how reluctant an object is to speed up. Rotation is trickier: a long wrench is easy to spin about its length but hard to spin end-over-end, so a single number is not enough. The inertia tensor packages all those direction-dependent stubbornnesses together, usually as a 3-by-3 grid of values.

The simpler idea hiding inside it is the moment of inertia: a measure of rotational reluctance about one chosen axis, set not just by how much mass a body has but by how that mass is spread out. Mass far from the spin axis counts much more than mass near it — this is why a figure skater spins faster by pulling their arms in, drawing mass toward the axis and lowering their moment of inertia. The full tensor simply records this reluctance for every direction at once, including the off-axis cross-terms (the products of inertia) that describe bodies whose mass is lopsided rather than neatly symmetric.

In robotics, every link of an arm — every forearm, every gripper segment — has its own inertia tensor, and these feed directly into the equations that predict how the robot will respond to the forces driving it. Get them wrong and the controller mispredicts the motion, leading to sloppy tracking or instability. That is why engineers carefully estimate each part's mass distribution, either from the 3D design model or by measurement, before trusting a robot to move quickly and precisely.